Plankton Atlas of the North Atlantic Ocean - Inter Research

MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser

Supplement: 3–10

Published 2004

Continuous Plankton Records: Plankton Atlas of the North Atlantic Ocean (1958–1999). I. Introduction and methodology Grégory Beaugrand1, 2,* 1

Sir Alister Hardy Foundation for Ocean Science, The Laboratory, Citadel Hill, Plymouth PL1 2PB, UK 2 CNRS, UMR 8013 ELICO, Station Marine, Université des Sciences et Technologies de Lille 1, 28 avenue Foch, BP 80, 62930 Wimereux, France

ABSTRACT: The Continuous Plankton Recorder (CPR) survey is one of the largest plankton monitoring programmes in the world. Data from this programme have been used to investigate many ecological issues such as biogeography, biodiversity and relationships between plankton and global change. A CPR Atlas on the geographical distribution of plankton in the North Atlantic Ocean was first published by the Edinburgh Oceanographic Laboratory (1973). This atlas, based on about 40 000 CPR samples collected from 1958 to 1968, contributed greatly to our knowledge of the biogeography of ~260 species or taxa in the North Atlantic Ocean. The data of the CPR Atlas were updated in the last 30 years. In this introduction to the revised CPR Atlas, which is based on data for 1958–1999, new numerical procedures and features used in preparing the updated maps are described. KEY WORDS: Continuous Plankton Recorder survey · Atlas · Plankton · North Atlantic Ocean Resale or republication not permitted without written consent of the publisher

The Continuous Plankton Recorder (CPR) survey is one of the largest plankton monitoring programmes in the world. Results from the survey have been used to investigate many ecological issues (Reid et al. 2003). Biogeographical studies have been conducted showing spatial distribution of more than 250 species throughout the North Atlantic Ocean and its shelf seas (Colebrook et al. 1961a,b, Edinburgh Oceanographic Laboratory 1973). Recently, the mapping protocol has been improved using the Lambert projection (Planque & Fromentin 1996) and mapping techniques such as kriging (Planque & Fromentin 1996, Planque et al. 1997). A number of papers have allowed a better characterisation of seasonal cycles and spatial changes for many taxa (Glover 1957, Colebrook 1979, 1984, 1991). Other studies have examined long-term changes in phytoplankton and zooplankton (Colebrook 1981, Reid et al. 1998, Beaugrand & Reid 2003). Results show that long-term variability in standing stock, production and community structure of the plankton might be related to the North Atlantic Oscillation (NAO) (Fromentin &

Planque 1996, Reid & Planque 2000). Other studies have recently shown a major reorganisation in the biodiversity of calanoid copepods related to regional warming (Beaugrand et al. 2002, Beaugrand 2003). Some studies have focussed on diel vertical migration of calanoid copepods (Hays et al. 1994, Hays 1995, 1996, Hirst & Batten 1998), spatial and temporal changes in the diversity of decapod crustacean larvae or calanoid copepods (Lindley 1998, Beaugrand et al. 2000, Beaugrand 2001). Monitoring of non-indigenous species (Edwards et al. 2001) and unusual events (Lindley et al. 1990, 1993, Edwards et al. 1999) have been undertaken as well. The CPR data have led to a better understanding of the ecology and functioning of North Atlantic ecosystems. A CPR Atlas on the geographical distribution of plankton in about 40 000 CPR samples collected from 1958 to 1968 from the North Atlantic Ocean was published by the Edinburgh Oceanographic Laboratory (1973). This Atlas contributed substantially to our knowledge of the biogeography of 255 species or taxa in the North Atlantic Ocean. However, no update of this monograph has been published in the last 30 years.

*Email: gregory.beaugrand@univ-lille1.fr

© Inter-Research 2004 · www.int-res.com

INTRODUCTION

4

Mar Ecol Prog Ser Suppl 2004: 3–10

Most of the maps presented in 1973 have been recreated in the new CPR Atlas, which summarises the work of the Continuous Plankton Recorder team (CPR Survey Team 2004; this issue), based on 155 669 samples collected from 1958 to 1999. Here in Part I, I briefly describe the methodology of collection and analysis of CPR samples and outline the new numerical procedure employed. New symbols and features have been added to the maps to increase the information and improve the interpretation.

CONTINUOUS PLANKTON RECORDER SURVEY The CPR survey is an upper layer plankton monitoring programme that has regularly collected samples in the North Atlantic and North Sea at monthly intervals since 1946 (Warner & Hays 1994). The CPR was first used during the RV ‘Discovery’ expedition to the Antarctic Ocean in 1925 to 1927. From 1931 it was regularly deployed along certain routes in the North Sea. The original idea was to use a similar methodology to that used in meteorological research to investigate causes and effects of changes in the abundance of marine plankton and to relate them to varying hydroclimatic conditions and catches of pelagic fishes such as herring (Hardy 1939). As the number of sampling years increased, it became possible to study changes in the abundance and composition of species through time. Since the start of the programme, data on the abundance of more than 400 species or taxa have been gathered by about 178 000 CPR samples collected up to the year 2000, which represents ~2 million entries and ~80 million data-points (Beaugrand et al. 2003).

Fig. 1. Geographical location of samples collected by the CPR survey from 1958 to 1999

The spatial distribution of CPR samples collected from 1958 to 1999 is presented in Fig. 1. The CPR programme is operated by a high-speed plankton recorder that is towed behind merchant ships at an average speed of 20 km h–1 and a depth of approximately 6.5 m (Hays & Warner 1993). Despite the limitation to near-surface sampling, studies have shown that this device gives a satisfactory assessment of the epipelagic zone (Lindley & Williams 1980, Batten et al. 1999). Water enters the recorder through a square aperture of 1.62 cm2 and flows to an area with a cross-section of 5 × 10 cm, where plankton is filtered by a slowly moving band of silk with an average mesh size of 270 µm. A second band of silk covers the organisms to form a sandwich that is reeled into a tank containing 4% formaldehyde. The speed of silk movement is adjusted to the speed of the ship by means of an impeller and a gearbox located above the internal mechanism; 1 cm on the silk corresponds to 1 nautical (n mile) mile of tow. At the laboratory, the silk roll is unwound and cut into sections corresponding to 10 n miles (18.5 km) or approximately 3 m3 of seawater filtered (Warner & Hays 1994). Positions and times of samples are calculated from data on the start and end of deployment as well as changes in the course and speed of the ship. Plankton identification and counting is realised in 4 steps: Step 1. Estimation of the colour of the silk into 4 categories: nil, very pale green, pale green, and green. This gives an index of chlorophyll concentration, also called the ‘greenness index’. This protocol has not changed since 1958. Step 2. Identification and quantification of phytoplankton taxa, to species level where possible. More than 200 phytoplankton species or taxa are identified. Subsampling of the filtering silk is conducted by examining 20 fields at 450 magnification (295 µm diameter view) in 2 diagonals of 10 fields across the silk. This corresponds to a subsample of about 1/8000 of the silk. Abundance of each phytoplankton taxonomic category is determined by counting the number of fields in which each taxon is detected. The methodology of this analysis has remained unchanged since 1958. Step 3. Examination of zooplankton, generally 2 mm on both filtering and covering silks. More than 150 species or taxa may be identified at this step. For counting zooplankton, a category system similar to a logarithmic progression of abundance is used to reduce the time of analysis. Methods of counting and data processing are described in more detail in Colebrook (1960, 1975), Warner & Hays (1994), Batten et al. (2003) and Reid et al. (2003).

5

Beaugrand: CPR Atlas methodology

NUMERICAL PROCEDURE FOR MAPPING SPATIAL DISTRIBUTIONS The new CPR Atlas is based on 155 669 CPR samples collected from 1958 to 1999 (see Fig. 1). Results for 240 species or taxonomic groups are presented in CPR Survey Team (2004, this volume). Other species are not included, as the data were often limited to a few records or in some cases to a single occurrence. The 1973 CPR Atlas presented the geographical distribution of 255 species or taxa. The numerical procedure used for the present CPR Atlas was divided into 2 main stages (Fig. 2). Stage 1. Spatial regularisation of species for each two-month and day/night period

Stage 1: Spatial regularisation of species for each 2-month and day/night period. This stage consists of 3 steps: (1) transformation of latitudes and longitudes into Lambert coordinates, (2) spatial regularisation and (3) mapping (Fig. 2). Step 1. Transformation of latitudes and longitudes into Lambert coordinates: In the 1973 CPR Atlas, the Mercator projection was used to map the geographical distribution of species. This projection, however, is not suitable for the mapping of large regions far away from the equator. For this updated version of the CPR Atlas, the Lambert Conic Conformal Projection was used (Planque & Fromentin 1996, Beaugrand et al. 2000, Stage 2. Calculation and mapping of the average spatial distribution of species

Step 1. Transformation of latitudes and longitudes into Lambert coordinates Time periods (12)

Latitude x Longitude (1066 100x100 n mile pixels)

Step 2: Spatial regularisation Latitude (26 pixels)

C. finmarchicus

Averaging

November-December night samples

Latitude x Longitude (1066 100x100 n mile pixels) 1

eri

od

s

Longitude (41 pixels)

tim

ep

C. finmarchicus

Latitude (26 pixels)

C. finmarchicus

12

January-February day samples

Reshaping

Longitude (41 pixels) Latitude

C. finmarchicus average spatial distribution

Longitude

Step 3: Mapping

Mapping

January-February (day data) 20°W 0°E

2.0

40 °N

0.5

30° N

1.573 1.049 40 ° 30 °

0.0

2.096

N

0.525

Abundance (in log10(x+1))

1.0

20°W 0°E


1.5

N

0.001

Fig. 2. Analysis stages and steps towards the production of the species distribution maps

6


2001), to avoid a strong deformation of distance on the maps. Step 2. Spatial regularisation: Sampling by the CPR is irregular in space, as the samples follow the tracks of ‘ships of opportunity’. Fig. 1 shows the spatial heterogeneity of CPR sampling. A regular grid (70° W to 20° E; 30 to 70° N) was defined on the basis of Lambert coordinates. The size of the geographical square for the spatial regularisation was fixed by trial and error to 100 × 100 n miles. This represents the best compromise between the number of missing geographical squares and the size of the spatial resolution. No spatial interpolation was used in the biogeographical charts. The spatial resolution was slightly improved in comparison to the 1973 CPR Atlas (geographical rectangle 1° latitude × 2° longitude). Spatial regularisation needs to be realised for homogeneous time periods. In the pelagic realm, diel and seasonal scales represent a source of variability often greater than at the interannual scale (van der Spoel 1994, Piontkovski et al. 1999, Beaugrand et al. 2003). Therefore, a numerical procedure was included to account for temporal scales of variability; 12 spatial regularisations were produced for each 2-month period (January-February; March-April; May-June;

July-August; September-October; November-December) for both day and night periods using the function in Beaugrand et al. (2001) (see their Fig. 1). The procedure had to address the problem of including both abundance and presence data, and it distinguished 3 cases: (1) species with only presence/ absence data; (2) species with only abundance data; (3) species with both presence and abundance data. Fig. 3 summarises the procedure implemented for each geographical square and time period. An average value of abundance was only calculated when the number of abundance data was > 2. A distinction was made between 1 and only 1 record (abundance or presence; symbolised by a ‘f’) and >1 record (abundance or presence; symbolised by a ‘+’). This was to differentiate the exceptional occurrence of a species in a region, possibly due to misidentification, contamination of a sample, miscalculation of the location of a sample, or a real occurrence related to atypical hydroclimatic forcing. When no abundance was calculated and the occurrence of a species was not detected for a pixel, a zero for no occurrence (symbolised by a ‘S’) was added on the map if the number of samples in the pixel was > 2 (fixed by trial and error). This was to avoid indications of no occurrence related to regions

CPR samples in the spatio-temporal window (e.g. In one 100 x 100 n mile pixel for the period January-February at night) Strictly superior to threshold 1

Abundance data

>1 non-zero observations

Only 1 non-zero observation

Abundance & presence data

Inferior or equal to threshold 1

Presence data

0 non-zero observations

-9999 Absence

Mean

-5555

-2222 >1 non-zero observations

Only 1 non-zero observation

mean

-9999

0 non-zero observations

Presence data=1

threshold 1=2

Only 1 presence

> 1 presence

-3333

-5555

-5555

-9999

blank

Presence data>1 -9999

Fig. 3. Procedure of spatial regularisation of species distributions. A ‘+’ (–9999): more than one occurrence in a geographical square for a given time period. R (–5555): one and only one occurrence in a geographical square. S (–3333): no occurrence in a geographical square. Blank (–2222): no occurrence in an undersampled region (≤ 2 samples for a given time period); these data are not distinguished from unsampled regions

7

Beaugrand: CPR Atlas methodology

January-February

30° N

20°W 0°E

40° N 30° N

September-October

March-April

30° N

20°W 0°E

40° N 30° N

November-December

May-June

20°W 0°E

40° N 30° N


30° N


20°W 0°E

40° N


40° N


20°W 0°E


40° N

July-August Abundance (in log10(x+1))

20°W 0°E

Fig. 4. Calanus glacialis. Seasonal changes in geographical distribution (nighttime data)

that were poorly sampled. No distinction was made between the latter regions and regions never sampled by the CPR, to avoid too much complexity in the maps. Step 3. Mapping: Although not shown in the CPR Atlas, mapping of the geographical distribution of each species for each of the 12 time periods is possible. Fig. 4 displays seasonal changes in the abundance of Calanus glacialis. This species was only common in the near surface water of the Labrador Current from January to June, emphasising the importance of seasonal variability in the spatial regulation. Diel vertical migration should also be taken into consideration (Fig. 5). For example, Pleuromamma abdominalis is mainly detected in near-surface waters during dark periods. Abundance categories were transformed using Colebrook’s (1975) function. Stage 2: Calculation and mapping of the average spatial distribution of species. Fig. 6 summarises the

procedure used to average the information (abundance and presence/absence) from the 12 maps. Thresholds used in that procedure were fixed to 6 time periods by trial and error. An average estimation of the abundance of a species for a geographical square was calculated when the number of 2-month time periods for which data on abundance were available was > 6 (out of a possible 12; 6 day and 6 night). An indication of presence (+) was added when a taxon was detected in > 2 samples (all time periods considered). An indication of an absence (S) was added in a geographical square when the absence of a taxon was confirmed in > 6 time periods (thresholds 2 and 3) and presence was never recorded. This ensured that absence of a species was reported in a relatively well-sampled region. Fig. 7 shows an example of maps produced by the procedure for the calanoid copepod Calanus finmarchicus. Table 1 summarises the main differences between the 1973 CPR Atlas and the present one.

8


b. Night distribution

a. Day distribution

0.50 40° N

0.25

30° N

0.75 0.50 40° N

0.25

30° N


0.75

1.00

20°W 0°E


1.00

20°W 0°E

0.00

0.00

Fig. 5. Pleurommama abdominalis. Geographical distribution (a) during the daytime and (b) at night for September-October

Elimination of negative data and recodage of -3333 in 0

Determination of the number of abundance data for each square

Number of abundance data < threshold 1

Number of abundance data >=threshold 1 Mean

All abundance data=0 or no abundance data

Mean=0

Mean>0

Abundance data>0 -9999

Research of presence and absence data

Mean

-5555 data=1

-5555

-5555 data>1

-9999 data>0

-9999

-9999

If -9999 data = 0 and -5555 data = 0 -3333

threshold 1=6 threshold 2=6 threshold 3=6

Research of presence and absence data

-5555 data=1 -5555

-5555 data>1

-9999 data>0

-9999

-9999

-3333 data

>=threshold 2 -3333

-2222 data

=threshold 3 -3333

1 presence data

Maximal value of abundance (in log10 (x+1))

Abundance category

20°W 0°E

Abundance coded in four shades of grey

2.096

Isobath 4000 m 1.573

1.049

Isobath 200 m 40

Blank: no sampling

°N

0.525 30

°N

0.001

Fig. 7. Calanus finmarchicus. Spatial distribution. Arrows indicate the main features on the map

Blank: no presence data recorded and sampling too weak to include a ‘S’

R : only 1 presence detected

Table 1. Comparison between the 1973 CPR Atlas and the present Atlas 1973 CPR Atlas Present Atlas Number of taxa Time period Number of data Projection

255 1958–1968 40 000 Mercator

Resolution Abundance category No. of symbols per map Variability

No

Other features

S: species never recorded

240 1958–1999 155 669 Lambert Conic Conformal 1° lat. × 2° long. 100 × 100 n miles Up to 3 Up to 4 Up to 4 Up to 7

Land mask

Seasonal and diel variabilitya Land mask and bathymetry

a

Available at: www.sahfos.org/CPR_atlas.htm

Acknowledgements. The authors are grateful to all past and present members and supporters of the Sir Alister Hardy Foundation for Ocean Science and its predecessors; their sustained help has allowed the establishment and maintenance of the long-term CPR dataset. The main support for this work was from the Department of Environment, Food and Rural Affairs (United Kingdom). Consortium support for the CPR Survey is provided by agencies from the following countries: United Kingdom, USA, Canada, Faroe Islands, France, Ireland, Netherlands, Portugal, and by the International Oceanographic Commission (IOC) and the European Union. We thank the owners, masters and crews of the ships that tow CPRs on a voluntary basis.

Minimal value of abundance (in log 10 (x+1))

LITERATURE CITED Batten SD, Hirst AG, Hunter J, Lampitt RS (1999) Mesozooplankton biomass in the Celtic Sea: a first approach to comparing and combining CPR and LHPR data. J Mar Biol Assoc UK 79:179–181 Batten SD, Clark RA, Flinkman J, Hays GC, and 6 others (2003) CPR sampling: the technical background, materials, and methods, consistency and comparability. Prog Oceanogr 58:193–215 Beaugrand G (2001) North Atlantic pelagic biodiversity and hydro-meteorological variability. PhD thesis, University Paris VI (Pierre et Marie Curie) Beaugrand G (2003) Long-term changes in copepod abundance and diversity in the north-east Atlantic in relation to fluctuations in the hydro-climatic environment. Fish Oceanogr 12:270–283 Beaugrand G, Reid PC (2003) Long-term changes in phytoplankton, zooplankton and salmon linked to climate change. Global Change Biol 9:801–817 Beaugrand G, Reid PC, Ibanez F, Planque B (2000) Biodiversity of North Atlantic and North Sea calanoid copepods. Mar Ecol Prog Ser 204:299–303 Beaugrand G, Ibañez F, Lindley JA (2001) Geographical distribution and seasonal and diel changes in the diversity of calanoid copepods of the North Atlantic and North Sea. Mar Ecol Prog Ser 219:189–203 Beaugrand G, Reid PC, Ibanez F, Lindley JA, Edwards M (2002) Reorganisation of North Atlantic marine copepod biodiversity and climate. Science 296:1692–1694 Beaugrand G, Ibañez F, Lindley JA (2003) An overview of statistical methods applied to the CPR data. Prog Oceanogr 58:235–262 Colebrook JM (1960) Continuous Plankton Records: methods of analysis, 1950–59. Bull Mar Ecol 41:51–54 Colebrook JM (1975) The Continuous Plankton Recorder survey: automatic data processing methods. Bull Mar Ecol 8:123–142

10


Colebrook JM (1979) Continuous Plankton Records: seasonal cycles of phytoplankton and copepods in the North Atlantic Ocean and the North Sea. Mar Biol 51:23–32 Colebrook JM (1981) Continuous Plankton Records: persistence in time-series of annual means of abundance of zooplankton. Mar Biol 61:143–149 Colebrook JM (1984) Continuous Plankton Records: relationships between species of phytoplankton and zooplankton in the seasonal cycle. Mar Biol 83:313–323 Colebrook JM (1991) Continuous Plankton Records: from seasons to decades in the plankton of the North-East Atlantic. In: Kawasaki T, Tanaka S, Toba Y, Taniguchi A (eds) Long-term variability of pelagic fish populations and their environments. Pergamon Press, Oxford, p 29–45 Colebrook JM, Glover RS, Robinson GA (1961a) Contribution towards a plankton atlas of the North-Eastern Atlantic and the North Sea. General introduction. Bull Mar Ecol 5:67–80 Colebrook JM, John DE, Brown WW (1961b) Contribution towards a plankton atlas of the North-Eastern Atlantic and the North Sea. Part II: Copepoda. Bull Mar Ecol 5: 90–97 CPR (Continuous Plankton Recorder) Survey Team (2004) Continuous Plankton Records: Plankton Atlas of the North Atlantic Ocean (1958–1999). II. Biogeographical charts. Mar Ecol Prog Ser Suppl 2004:11–75 Edinburgh Oceanographic Laboratory (1973) Continuous Plankton Records: a plankton atlas of the North Atlantic and the North Sea. Bull Mar Ecol 7:1–174 Edwards M, John AWG, Hunt HG, Lindley JA (1999) Exceptional influx of oceanic species into the North Sea late 1997. J Mar Biol Ass UK 79:737–739 Edwards M, John AWG, Johns DG, Reid PC (2001) Case history and persistence of the non-indigenous diatom Coscinodiscus wailesii in the north-east Atlantic. J Mar Biol Assoc UK 81:207–211 Fromentin JM, Planque B (1996) Calanus and environment in the eastern North Atlantic. II. Influence of the North Atlantic Oscillation on C. finmarchicus and C. helgolandicus. Mar Ecol Prog Ser 134:111–118 Glover RS (1957) An ecological survey of the Scottish herring fishery. Part II: the planktonic environment of the herring. Bull Mar Ecol 5:1–43 Hardy AC (1939) Ecological investigations with the Continuous Plankton Recorder: object, plan, methods. Hull Bull Mar Ecol 1:1–57 Hays GC (1995) Ontogenic and seasonal variation in the diel vertical migration of the copepods Metridia lucens and Metridia longa. Limnol Oceanogr 40:1461–1465 Hays GC (1996) Large-scale patterns of diel vertical migration in the North Atlantic. Deep-Sea Res I 43:1601–1615

Hays GC, Warner AJ (1993) Consistency of towing speed and sampling depth for the Continuous Plankton Recorder. J Mar Biol Assoc UK 73:967-970 Hays GC, Proctor CA, John AWG, Warner AJ (1994) Interspecific differences in the diel vertical migration of marine copepods: the implications of size, color, and morphology. Limnol Oceanogr 39:1621–1629 Hirst AG, Batten SD (1998) Long-term changes in the diel vertical migration behaviour of Calanus finmarchicus in the North Sea are unrelated to fish predation. Mar Ecol Prog Ser 171:307–310 Lindley JA (1998) Diversity, biomass and production of decapod crustacean larvae in a changing environment. Invertebr Reprod Dev 33:209–219 Lindley JA, Williams R (1980) Plankton of the Fladen Ground during FLEX 76 II. Population dynamics and production of Thysanoessa inermis (Crustacea: Euphausiacea). Mar Biol 57:79–86 Lindley JA, Roskell J, Warner AJ, Halliday NC, Hunt HG, John AWG, Jonas TD (1990) Doliolids in the German Bight in 1989: evidence for exceptional inflow into the North Sea. J Mar Biol Assoc UK 70:679–682 Lindley JA, Williams R, Hunt HG (1993) Anomalous seasonal cycles of decapod crustacean larvae in the North Sea plankton in an abnormally warm year. J Exp Mar Biol Ecol 173:47–65 Piontkovski S, van der Spoel S, Prusova I (1999) Diurnal rhythm of biodiversity in a zooplankton community of a macroscale anticyclonic gyre. Crustaceana 72:1–15 Planque B, Fromentin JM (1996) Calanus and environment in the eastern North Atlantic. I. Spatial and temporal patterns of C. finmarchicus and C. helgolandicus. Mar Ecol Prog Ser 134:101–109 Planque B, Hays GC, Ibanez F, Gamble JC (1997) Large scale spatial variations in the seasonal abundance of Calanus finmarchicus. Deep-Sea Res I 44:315–326 Reid PC, Planque B (2000) Long-term planktonic variations and the climate of the North Atlantic. In: Mills D (ed) The ocean life of Atlantic salmon: environmental and biological factors influencing survival. Fishing News Books, London, p 153–169 Reid PC, Edwards M, Hunt HG, Warner AJ (1998) Phytoplankton change in the North Atlantic. Nature 391:546 Reid PC, Colebrook JM, Matthews JBL, Barnard R and 24 others (2003) The Continuous Plankton Recorder: concepts and history, from plankton indicator to undulating recorders. Prog Oceanogr 58:117–173 van der Spoel S (1994) History, progress and future of theory in pelagic biogeography. Prog Oceanogr 34: 101–107 Warner AJ, Hays GC (1994) Sampling by the Continuous Plankton Recorder survey. Prog Oceanogr 34:237–256